CN110206743B - Axial flow pump cavitation feature extraction method based on noise texture and bubble state contrast - Google Patents
Axial flow pump cavitation feature extraction method based on noise texture and bubble state contrast Download PDFInfo
- Publication number
- CN110206743B CN110206743B CN201910450974.2A CN201910450974A CN110206743B CN 110206743 B CN110206743 B CN 110206743B CN 201910450974 A CN201910450974 A CN 201910450974A CN 110206743 B CN110206743 B CN 110206743B
- Authority
- CN
- China
- Prior art keywords
- image
- flow pump
- axial flow
- cavitation
- noise
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 19
- 238000004458 analytical method Methods 0.000 claims abstract description 19
- 238000001914 filtration Methods 0.000 claims abstract description 17
- 230000009467 reduction Effects 0.000 claims abstract description 10
- 238000012795 verification Methods 0.000 claims abstract description 7
- 238000000034 method Methods 0.000 claims description 28
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 238000011161 development Methods 0.000 claims description 9
- 230000000737 periodic effect Effects 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 8
- 238000011049 filling Methods 0.000 claims description 5
- 230000000877 morphologic effect Effects 0.000 claims description 5
- 230000001133 acceleration Effects 0.000 claims description 3
- 238000000926 separation method Methods 0.000 claims description 2
- 210000003934 vacuole Anatomy 0.000 claims 2
- 238000012545 processing Methods 0.000 abstract description 6
- 238000010586 diagram Methods 0.000 description 11
- 238000011160 research Methods 0.000 description 6
- 239000000284 extract Substances 0.000 description 4
- 239000012530 fluid Substances 0.000 description 4
- 230000000977 initiatory effect Effects 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000012512 characterization method Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000009432 framing Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 238000007792 addition Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000013525 hypothesis research Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000011426 transformation method Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Images
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F04—POSITIVE - DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS FOR LIQUIDS OR ELASTIC FLUIDS
- F04D—NON-POSITIVE-DISPLACEMENT PUMPS
- F04D15/00—Control, e.g. regulation, of pumps, pumping installations or systems
- F04D15/0088—Testing machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Image Processing (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses an axial flow pump cavitation feature extraction method based on noise texture and bubble state contrast, which comprises the following steps: (1) collecting axial flow pump video data with cavitation bubbles as an experimental group, collecting axial flow pump video data without cavitation under the same working condition as a control group, corresponding each frame image of the working state of the axial flow pump in the two groups of video data, and eliminating a background image; (2) converting image frames of two groups of video data into gray level images, extracting edges of the images and then carrying out noise reduction processing; (3) obtaining image signal characteristics of bubbles in the experimental group according to the characteristic information of the two groups of images; (4) filtering and denoising the acoustic signal data of the experimental group, then carrying out frequency domain analysis, and extracting noise texture characteristics; (5) and comparing the noise texture characteristics with the image signal characteristics of the bubbles and establishing a relation between the noise texture characteristics and the image signal characteristics of the bubbles. The invention realizes the purpose of extracting the cavitation feature of the axial flow pump by mutual verification of the acoustic signal feature and the image feature.
Description
Technical Field
The invention belongs to the field of axial flow pump signal processing, and particularly relates to an axial flow pump cavitation feature extraction method based on noise texture and bubble state comparison.
Background
The fluid machinery comprises a pump, a hydrofoil, an axial flow pump and the like, the cavitation and cavitation phenomena are often generated when the fluid machinery runs under the actual flow field working condition, the processes of cavitation initiation, growth development, collapse and the like are continuously generated due to the change of air pressure and water pressure in the flow field, particularly impact energy generated by bubble collapse in the collapse process is corroded on the fluid machinery under the long-term accumulation effect, and the reduction of the mechanical performance is difficult to avoid. Meanwhile, the same fluid machine (such as an axial flow pump) can generate different cavitation development conditions under different working conditions, such as cavitation initiation, tip vortex cavitation, sheet cavitation, cloud cavitation and the like. When cavitation develops to a certain extent, it will cause great damage to mechanical properties and bring about noise. Therefore, for engineering application, the method has important practical significance for timely finding the generation of cavitation and preventing further development of the cavitation, and needs a learner to deeply research the cavitation mechanism and summarize the cavitation characteristics.
For example, chinese patent publication No. CN102252748A discloses a method for extracting modulation characteristics of cavitation noise based on empirical mode, which utilizes the adaptivity of empirical mode decomposition and the high resolution of Hilbert-Huang transform, and overcomes the disadvantage that the conventional modulation characteristic extraction method is difficult to extract modulation characteristics of short-time non-stationary modulated cavitation noise data.
At present, a certain achievement and theoretical basis are obtained by analyzing the characteristics of cavitation purely by using acoustic signals, but the mechanistic and academic circles of cavitation lack unified understanding. Meanwhile, the deep level of research on cavitation is not matched with the wide application prospect in the current research direction, so that a means for helping scholars to establish a reliable cavitation characteristic research method is urgently needed.
Disclosure of Invention
The invention provides an axial flow pump cavitation feature extraction method based on noise texture and bubble state contrast, which realizes the purpose of detecting cavitation bubbles of an axial flow pump by an acoustic signal through mutual verification of acoustic signal features and image features.
The technical scheme of the invention is as follows:
an axial flow pump cavitation feature extraction method based on noise texture and bubble state contrast comprises the following steps:
(1) acquiring axial flow pump video data and acoustic signal data with cavitation bubbles as experimental groups, acquiring axial flow pump video data and acoustic signal data without cavitation bubbles under the same working condition as a control group, and corresponding each frame of video image of the working state of the axial flow pump in the two groups of video data;
(2) converting video image frames of the two groups of video data into gray level images, extracting the edges of the images, and then carrying out noise reduction treatment;
(3) obtaining image signal characteristics of bubbles in the experimental group according to the characteristic information of the two groups of images;
(4) carrying out background cancellation and filtering on the acoustic signal data of the experimental group, carrying out frequency domain analysis, and extracting noise texture characteristics (bubble area, quantity, change rate, diameter and the like);
(5) and comparing the noise texture features with the image signal features of the bubbles, and establishing a relation between the noise texture features and the image signal features according to the matching property in a time domain, the relation between the main frequency of the image bubble area changing along with time and the main frequency of the acoustic signal, the relation between the bubble movement acceleration and the acoustic signal intensity and the like.
The method realizes the characteristic extraction and characterization of the cavitation bubbles of the axial flow pump in the actual flow field work, extracts the parameter characteristics capable of reflecting the periodic behavior development of the cavitation bubbles from the time domain, and extracts the parameter characteristics capable of reflecting the morphological evolution of the cavitation bubbles from the frequency spectrum. By utilizing the periodicity of the axial flow pump, the periodic characteristics of cavitation changing along with time, such as cavitation initiation, development, collapse and the like, can be obtained and compared with the characteristics of acoustic signals, so that the acoustic texture information of cavitation is printed.
In the step (2), the specific process of extracting the edge of the image is as follows: the edge of the image is extracted by using an edge extraction method, the background noise of the axial flow pump is converted into linear noise, the influence of the background of the axial flow pump on bubbles is eliminated, and the separation of target cavitation bubble information and the background of the axial flow pump is realized.
The noise reduction processing comprises the following steps: filtering the image to remove the point noise in the image; and carrying out binarization on the image, and filling the holes of the image.
In the step (3), the method for obtaining the image signal characteristics of the bubbles in the experimental group comprises the following steps: and counting the characteristic information of the two groups of images, and subtracting the two characteristic information to obtain the image signal characteristics of the bubbles in the experimental group.
In addition, in the step (3), the method for obtaining the image signal characteristics of the bubbles in the experimental group can also adopt a silhouette algorithm, take the image of the contrast group as a background noise signal, subtract the matrixes of the two groups of gray-scale images, remove the edge background of the axial-flow pump, and extract the image signal characteristics of the bubbles. With this method, more reliable results can be obtained.
In the step (4), the specific steps of background cancellation and filtering of the acoustic signal data of the experimental group are as follows:
using the collected acoustic signal data of the control group as background noise information, and performing background cancellation by using a recursive least square algorithm (RLS) or a least mean square algorithm (LMS) according to the background noise information and the actual acoustic signal data of the experimental group; and determining a filtering means according to the main cavitation frequency band range of the axial flow pump needing to be analyzed.
The signal noise reduction processing by using the background data can eliminate some inherent background signals, highlight the actually required cavitation noise for further analysis and reduce noise interference.
In the step (4), the frequency domain analysis method includes, but is not limited to, one of wavelet transform, fast fourier transform, short-time fourier transform, and Hilbert-Huang transform.
In the step (5), because the acoustic data and the image data are acquired simultaneously, the results obtained after the data processing of the acoustic data and the image data are in time domain connection, and the time domain characteristics on the acoustic signal can be corresponded to the specific image signal through the time domain, and are analyzed and verified; parameter characteristics capable of reflecting the periodic behavior development of the cavitation bubbles in the acoustic data time domain are searched, verified and analyzed through the image data, and then the parameter characteristics capable of reflecting the morphological evolution of the cavitation bubbles are extracted from the acoustic data according to the characteristics. And finally, the regression image is subjected to result verification, namely image characteristics-signal time domain parameter characteristics-signal frequency domain parameter characteristics-signal time domain parameter characteristics-image characteristics. The process of continuous comparison, analysis and verification is established.
The method adopts two groups of axial flow pump cavitation videos or images, and eliminates noise and obtains a binary outline image of bubbles by using image filtering and some transformation methods aiming at the characteristics of strong background noise and strong dynamic transient property of the axial flow pump images, thereby counting the image characteristics of the bubbles; and noise reduction is realized by using an acoustic background cancellation technology to extract acoustic signal characteristics with high parameter values and high resolution. Analyzing the distribution characteristics and differences of noise frequency bands for stable cavitation types such as tip vortex cavitation, sheet cavitation and supercavitation; and for unstable cavitation types such as cloud cavitation which periodically occur, the evolution of a noise frequency band in a single period and periodic texture characteristics caused by the periodic cavitation phenomenon are analyzed. Finally, the matching of the acoustic signal and the image signal in time is combined for evidence analysis, so that the theoretical analysis result is tested, and the research on cavitation forming mechanism and cavitation characteristic analysis is facilitated.
Drawings
FIG. 1 is a schematic flow chart of an axial flow pump cavitation feature extraction method based on noise texture and bubble morphology contrast according to the present invention;
FIG. 2 is a schematic diagram of gray scale image conversion and framing of target cavitation regions for images of an experimental group according to an embodiment of the present disclosure;
FIG. 3 is a diagram illustrating edge extraction performed on an image according to an embodiment of the present invention;
FIG. 4 is a diagram illustrating filtering an image according to an embodiment of the present invention;
FIG. 5 is a schematic diagram illustrating binarization and hole filling of an image according to an embodiment of the present invention;
FIG. 6 is a diagram of average area of cavitation bubbles versus time obtained after statistics of image characteristic parameters according to an embodiment of the present invention;
FIG. 7 is a time-frequency diagram obtained by a short-time Fourier transform method before background cancellation of acoustic signal data;
FIG. 8 is a time-frequency diagram obtained by short-time Fourier transform after background cancellation of acoustic signal data;
FIG. 9 is a time-frequency diagram obtained by wavelet transform after background cancellation and denoising of acoustic signal data according to an embodiment of the present invention;
fig. 10 is a time-frequency diagram obtained by the acoustic signal data after background cancellation and noise reduction by a short-time fourier transform method in the embodiment of the method.
Detailed Description
The invention will be described in further detail below with reference to the drawings and examples, which are intended to facilitate the understanding of the invention without limiting it in any way.
As shown in fig. 1, an axial flow pump cavitation feature extraction method based on noise texture and bubble morphology contrast includes the following steps:
s01, firstly, using the video with cavitation bubbles obtained by the high-speed camera as an experimental group, using the video without cavitation bubbles at the same rotating speed as a control group, opening two video images in the ImageJ software, and corresponding the image frame number under the corresponding working condition.
Because the experiment has the same exposure frequency when shooting videos by using a high-speed camera, the same frame frequency and playing speed are kept, and the condition that each frame of the obtained two videos corresponds to the same working condition can be ensured. In the AVI video file in the embodiment, the frequency of the axial flow pump is 21Hz, the video playing speed is reduced to 1/500, and the video is played for 30 frames per second. The axial flow pump is in a flowing water field, the actual working condition of the axial flow pump can be approximately simulated, and the influence of bubbles in a gravity field can be ignored. The plane of the video image is parallel to the axial line of the axial flow pump, so that when the axial flow pump works, the circumferential flowing of bubbles can be seen when the axial flow pump rotates, and the radial information cannot be obtained.
Because the axial flow pump is a revolving body, although only a cavitation image on a single blade at a certain moment can be obtained, the cavitation conditions on other blades can be expanded. The lamellar cavitation tends to adhere to the axial flow pump blade, and the form of the lamellar cavitation can still be observed on the two-dimensional image.
And S02, converting the video image frames of the two groups of video data into gray maps, extracting the edges of the images, and then carrying out noise reduction processing.
For the acquisition of the bubble image, the RGB image of three channels has no great statistical significance on bubble feature statistics, and the gray-scale image can convert the three channels in each pixel into a single channel, so that the calculated amount is reduced, and the efficiency is improved, therefore, the unit-8 operation is used for converting the RGB three-color image into the gray-scale image, each pixel of the image matrix is only provided with one sampling color for representing the depth, and the sampling color is used for laying a cushion for subsequent edge and filling.
Due to the rotation periodicity of the axial flow pump in the video and the fact that the obtained video image is only shot from one side face of the axial flow pump, the all cavitation images of one blade of the axial flow pump can be obtained at each moment, and the slice-shaped cavitation on other blades has an angle with the shooting face and the visible area is not equal to the actual area, so that the slice-shaped cavitation bubble image of the blade perpendicular to the horizontal plane in the video is only selected.
When the original blade is far away from the lens along with the rotation, the next blade just turns into the main lens, the lens captures the cavitation phenomenon which possibly appears on the next blade, the cavitation phenomenon is analyzed and counted, meanwhile, the periodicity of the axial flow pump is corresponded, and as shown in fig. 1, a schematic diagram of performing gray scale image conversion and framing a target cavitation area for an image is shown.
The images in this block are then used to extract the edge images of the bubbles using Find Edges, as shown in fig. 3. The images are filtered using Fast Filters in Plugins-Filters or other low pass Filters to remove point noise in the video, as shown in FIG. 4. Here, a PureDenoise plug-in may also be utilized, which is more targeted to point noise.
The method comprises the steps of selecting Brightness/Contrast in Image-Adjust, selecting a proper value to binarize a video Image, jumping out a bullet frame, and adjusting a max slide bar and a min slide bar in the bullet frame to determine a binarization threshold value, wherein generally, the max value is equal to the min value (max min a), namely, the interval range of the threshold value is reduced to a value, and then an Image with only black and white color can be extracted, the white color is a bubble outline, and the black color is a background. And filling Holes by using Fill Holes to obtain an actual image of the bubbles, wherein in the actual operation, the center of the bubbles is white during binarization due to a certain gray value, so that the step of Fill Holes can be omitted. And (4) the threshold value a of the binarization is ensured to be the same as much as possible, so that the background image in the control group is consistent with the background image in the experimental group. The resulting effect is shown in fig. 5.
And S03, counting the characteristic information of the two groups of images, and subtracting the two characteristic information to obtain the image signal characteristics of the bubbles in the experimental group.
And in statistics, Image-Adjust-Threshold is used for channel screening of the video Image, and then analysis-analysis partitions are used for statistics. Here, the desired statistic is selected as needed.
In this embodiment, MATLAB analysis is directly performed on the obtained Total Area and Count data, the data of the control group is subtracted from the data in the experimental group, the corresponding values of the control group are first subtracted from the Total Area and Count in the experimental group, the average Area of the bubbles is obtained from the average Area and the Total Area/Count, and thus actual bubble data is obtained, and an image is drawn for comparison with the acoustic image, as shown in fig. 6.
In the step S03, the control group may be input as a background signal and the experimental group may be input as a main signal by using a silhouette algorithm, so as to obtain an image with only a cavitation signal, and then the image is used in the subsequent steps.
And S04, performing background cancellation and filtering on the acoustic signal data of the experimental group, performing frequency domain analysis, and extracting noise texture features.
Firstly, carrying out background cancellation operation, importing acoustic data, recording the background noise data of a control group as d (n), recording the original data of an experimental group as X (n), and carrying out the following steps:
e(n)＝d(n)-y(n)
wi(n+1)＝wi(n)+2μe(n)x(n-i)
solving a matrix equation:
y(n)＝WT(n)X(n)
e(n)＝d(n)-y(n)
W(n+1)＝W(n)+2μe(n)X(n)
wherein W (n) ═ w0(n),w1(n),w2(n),w3(n)......wN-1(n)]TN is the order of the filter, e (N) is the iterative error, namely the output value after the background which is finally solved is cancelled; the time-frequency diagram of the original data is shown in FIG. 7, and the calculation results of the algorithm are shown in the followingFig. 8. The inherent frequency (background noise such as main shaft frequency) of the signal is better improved after the signal is subjected to background cancellation means, and the parameter characteristic of the periodic behavior development of the cavitation signal in the time domain is highlighted, but the parameter characteristic of the cavitation signal in the frequency domain, which can be used for reflecting the morphological evolution of cavitation bubbles, cannot be seen at the moment. This will be solved by filtering by the next means.
For most axial flow pump acoustic signal data, the frequency band of the cavitation bubbles exists mainly in the range of 0-500Hz (this range is clearly characteristic and does not represent existence in other frequency band ranges), while the distribution range of the background signal often reaches the order of magnitude of 104Above, therefore, low-pass filtering is required; the embodiment adopts an inverse short-time Fourier transform (ISTFT) means, the algorithm is high in operation speed and efficiency, and the reliability of the filtering applied to the axial flow pump is high after the algorithm is verified. The video images obtained after filtering are shown in fig. 9 and fig. 10, where fig. 9 is a time-frequency image obtained by wavelet transform means, and fig. 10 is a time-frequency image obtained by short-time fourier transform means. Now we get more signal analysis maps of the parameter characteristics, so that more analysis operations can be performed.
The method can extract acoustic features by utilizing wavelet, Fast Fourier Transform (FFT), Short Time Fourier Transform (STFT), Hilbert-yellow (Hilbert-Huang) transform and the like; the method is applied to the method because the cavitation acoustic signal analysis results of the axial flow pump (21HZ rotating speed) are ideal through experiments and practical inspection and analysis.
And S05, comparing the noise texture features with the image signal features of the bubbles, and establishing the relation between the noise texture features and the image signal features according to the matching property in the time domain, the relation between the main frequency of the image bubble area changing along with the time and the main frequency of the acoustic signal, and the relation between the bubble movement acceleration and the acoustic signal intensity.
Since the image data and the acoustic data have matching in the time domain, the analysis data of the above steps S03 and S04 are compared and verified with the cavitation state image and the acoustic data at the same time. Parameter characteristics capable of reflecting the periodic behavior development of the cavitation bubbles in the acoustic data time domain are searched, verified and analyzed through the image data, and then the parameter characteristics capable of reflecting the morphological evolution of the cavitation bubbles are extracted from the acoustic data according to the characteristics. And finally, the regression image is subjected to result verification, namely image characteristics-signal time domain parameter characteristics-signal frequency domain parameter characteristics-signal time domain parameter characteristics-image characteristics. The process of continuous comparison, analysis and verification is established.
The method finally aims to achieve an efficient, reliable and accurate axial-flow pump cavitation characterization method through the working process established in the step S05, establish a complete axial-flow pump cavitation analysis system, provide a method reference for theoretical basic research for developing axial-flow pump cavitation state monitoring, provide a process means for deeply understanding an axial-flow pump cavitation mechanism, and provide a technical means for verifying the axial-flow pump cavitation characteristic research hypothesis.
The embodiments described above are intended to illustrate the technical solutions and advantages of the present invention, and it should be understood that the above-mentioned embodiments are only specific embodiments of the present invention, and are not intended to limit the present invention, and any modifications, additions and equivalents made within the scope of the principles of the present invention should be included in the scope of the present invention.
Claims (6)
1. An axial flow pump cavitation feature extraction method based on noise texture and bubble state contrast is characterized by comprising the following steps:
(1) acquiring axial flow pump video data and acoustic signal data with cavitation bubbles as experimental groups, acquiring axial flow pump video data and acoustic signal data without cavitation bubbles under the same working condition as a control group, and corresponding each frame of video image of the working state of the axial flow pump in the two groups of video data;
(2) converting video image frames of the two groups of video data into gray level images, extracting the edges of the images, and then carrying out noise reduction treatment;
(3) obtaining image signal characteristics of bubbles in the experimental group according to the characteristic information of the two groups of images;
(4) carrying out background cancellation and filtering on the acoustic signal data of the experimental group, carrying out frequency domain analysis, and extracting noise texture characteristics;
(5) comparing the noise texture features with the image signal features of the bubbles, and establishing a relation between the noise texture features and the image signal features of the bubbles according to the matching property in a time domain, the relation between the main frequency of the image bubble area changing along with time and the main frequency of the acoustic signal, and the relation between the bubble movement acceleration and the acoustic signal intensity; the method specifically comprises the following steps:
parameter characteristics capable of reflecting the periodic behavior development of the vacuoles in the acoustic data time domain are searched, verified and analyzed through image data, then the parameter characteristics capable of reflecting the morphological evolution of the vacuoles are extracted from the acoustic data according to the characteristics, and finally the image is regressed to carry out result verification.
2. The axial-flow pump cavitation feature extraction method based on noise texture and bubble morphology contrast as claimed in claim 1, wherein in step (2), the specific process of extracting the edge of the image is as follows:
the edge of the image is extracted by using an edge extraction method, the background noise of the axial flow pump is converted into linear noise, the influence of the background of the axial flow pump on bubbles is eliminated, and the separation of target cavitation bubble information and the background of the axial flow pump is realized.
3. The axial-flow pump cavitation feature extraction method based on noise texture and bubble morphology contrast as claimed in claim 1, wherein in step (2), the noise reduction treatment comprises:
filtering the image to remove the point noise in the image; and carrying out binarization on the image, and filling the holes of the image.
4. The axial flow pump cavitation feature extraction method based on noise texture and bubble morphology contrast as claimed in claim 1, wherein in step (3), the method for obtaining the image signal features of the bubbles in the experimental group is:
and counting the characteristic information of the two groups of images, and subtracting the two characteristic information to obtain the image signal characteristics of the bubbles in the experimental group.
5. The axial flow pump cavitation feature extraction method based on noise texture and bubble morphology contrast as claimed in claim 1, wherein in step (3), the method for obtaining the image signal features of the bubbles in the experimental group is:
and by a silhouette algorithm, taking the contrast group image as a background noise signal, subtracting the matrixes of the two groups of gray-scale images, removing the edge background of the axial-flow pump, and extracting the image signal characteristic of the air bubbles.
6. The axial flow pump cavitation feature extraction method based on noise texture and bubble morphology contrast as claimed in claim 1, wherein in step (4), the specific steps of background cancellation and filtering of acoustic signal data of the experimental group are as follows:
using the collected acoustic signal data of the control group as background noise information, and performing background cancellation by using a recursive least square algorithm or a least mean square algorithm according to the background noise information and the actual acoustic signal data of the experimental group; and determining a filtering means according to the main cavitation frequency band range of the axial flow pump to be analyzed, and filtering.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910450974.2A CN110206743B (en) | 2019-05-28 | 2019-05-28 | Axial flow pump cavitation feature extraction method based on noise texture and bubble state contrast |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910450974.2A CN110206743B (en) | 2019-05-28 | 2019-05-28 | Axial flow pump cavitation feature extraction method based on noise texture and bubble state contrast |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110206743A CN110206743A (en) | 2019-09-06 |
| CN110206743B true CN110206743B (en) | 2020-05-08 |
Family
ID=67789027
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910450974.2A Active CN110206743B (en) | 2019-05-28 | 2019-05-28 | Axial flow pump cavitation feature extraction method based on noise texture and bubble state contrast |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110206743B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114813037B (en) * | 2022-04-21 | 2023-06-20 | 中国船舶科学研究中心 | Cavitation flow structure frequency distribution characteristic analysis method |
| CN117409001B (en) * | 2023-12-14 | 2024-04-05 | 合肥晶合集成电路股份有限公司 | Bubble analysis method and analysis device for wafer bonding |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1847824B (en) * | 2006-02-27 | 2011-01-05 | 西安交通大学 | Analysis system and method of microbubble behavior in ultrasonic field based on superhigh speed photograph technology |
| JP2007303889A (en) * | 2006-05-09 | 2007-11-22 | Ho Jinyama | Cavitation detection method, evaluation method, computer program, and cavitation detection device |
| CN104535645B (en) * | 2014-12-27 | 2016-06-29 | 西安交通大学 | Microsecond differentiates the three-dimensional cavitation quantitative imaging method of cavitation spatial and temporal distributions |
| US20180087499A1 (en) * | 2016-09-23 | 2018-03-29 | Caterpillar Inc. | System for detecting faults in a pump |
| CN107273570A (en) * | 2017-05-10 | 2017-10-20 | 江苏大学 | A kind of blade pump cavitation Inductive noise Numerical Prediction Method |
| CN107956708B (en) * | 2017-11-17 | 2019-04-02 | 浙江大学 | A kind of potential cavitation fault detection method of pump based on quick spectrum kurtosis analysis |
-
2019
- 2019-05-28 CN CN201910450974.2A patent/CN110206743B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN110206743A (en) | 2019-09-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11443437B2 (en) | Vibe-based three-dimensional sonar point cloud image segmentation method | |
| CN109035149B (en) | License plate image motion blur removing method based on deep learning | |
| CN107220988B (en) | Part image edge extraction method based on improved canny operator | |
| CN110595780B (en) | Bearing fault identification method based on vibration gray level image and convolution neural network | |
| CN110206743B (en) | Axial flow pump cavitation feature extraction method based on noise texture and bubble state contrast | |
| CN108600135A (en) | A kind of recognition methods of signal modulation mode | |
| CN109472788B (en) | Method for detecting flaw on surface of airplane rivet | |
| CN107742307A (en) | Based on the transmission line galloping feature extraction and parameters analysis method for improving frame difference method | |
| CN104504662A (en) | Homomorphic filtering based image processing method and system | |
| CN103208097A (en) | Principal component analysis collaborative filtering method for image multi-direction morphological structure grouping | |
| CN107330891B (en) | Effective image quality evaluation system | |
| CN117094975A (en) | Method and device for detecting surface defects of steel and electronic equipment | |
| CN113899349B (en) | Sea wave parameter detection method, equipment and storage medium | |
| CN110531362B (en) | High-resolution motion sonar knowledge base target detection method | |
| CN114386191A (en) | Method for detecting gear engagement, random forest model, training method and system | |
| Jia | Fabric defect detection based on open source computer vision library OpenCV | |
| CN110427817B (en) | Hydrofoil cavitation feature extraction method based on cavitation image positioning and acoustic texture analysis | |
| CN112381845B (en) | Rock core image generation method, model training method and device | |
| CN109547777B (en) | Method for rapidly detecting video noise of complex scene | |
| CN112614109A (en) | Image quality evaluation method, device and computer readable storage medium | |
| CN107369163A (en) | A kind of quick SAR image object detection method based on best entropy Double Thresholding Segmentation | |
| CN111147804B (en) | Video frame reconstruction method | |
| CN113469913B (en) | Hot-rolled strip steel surface water drop removing method based on gradual cycle generation countermeasure network | |
| Cao et al. | Vehicle objects detection of video images based on gray-scale characteristics | |
| CN112561895A (en) | Airtightness leakage grade evaluation method and system based on artificial intelligence |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |